Predators may influence their prey populations not only through direct lethal effects, but also through indirect behavioral changes. Here, we combined spatiotemporal fine-scale data from GPS radio collars on lions with habitat use information on 11 African herbivores in Hwange National Park (Zimbabwe) to test whether the risk of predation by lions influenced the distribution of herbivores in the landscape. Effects of long-term risk of predation (likelihood of lion presence calculated over four months) and short-term risk of predation (actual presence of lions in the vicinity in the preceding 24 hours) were contrasted. The long-term risk of predation by lions appeared to influence the distributions of all browsers across the landscape, but not of grazers. This result strongly suggests that browsers and grazers, which face different ecological constraints, are influenced at different spatial and temporal scales in the variation of the risk of predation by lions. The results also show that all herbivores tend to use more open habitats preferentially when lions are in their vicinity, probably an effective anti-predator behavior against such an ambush predator. Behaviorally induced effects of lions may therefore contribute significantly to structuring African herbivore communities, and hence possibly their effects on savanna ecosystems.

No abstract

Summary1. An emerging perspective in the study of density dependence is the importance of the spatial and temporal heterogeneity of resources. Although this is well understood in temperate ungulates, few studies have been conducted in tropical environments where both food and water are limiting resources. 2. We studied the regulation of one of the world's largest elephant populations in Hwange National Park, Zimbabwe. The study period started in 1986 when the population was released from culling. Using census data we investigated changes in elephant abundance with respect to rainfall and density across the entire park and across waterholes. 3. The population more than doubled since culling stopped. The population increased continuously during the first 6 years, and then fluctuated widely at about 30 000 individuals. Immigration processes must have been involved in the increase of the population size. 4. Population growth rates were negatively related to previous population density by a convex relationship, and negatively related to the ratio of previous population density on annual rainfall by a linear relationship. However, only this latter model (i.e. assuming a fluctuating carrying capacity related to annual rainfall) produced realistic dynamics. Overall, population decreased during dry years when the elephant density was high. 5. During dry years there were fewer waterholes retaining water during the dry season and consequently elephant numbers at waterholes increased, while their aggregation level across waterholes decreased. On the long-run elephant numbers increased only at the less crowded waterholes. 6. We suggest that the interaction between population size and the available foraging range determined by the number of active waterholes during the dry season controls the park population. 7. Our results emphasize the need to understand how key-resource areas cause resource-based aggregation, which ultimately influences the strength of density dependence. More specifically, this study suggests that climate variability strongly affects local elephant population dynamics through changes in surface-water availability. Finally, as dispersal is likely to be an important driver of the dynamics of this population, our results support views that a metapopulation framework should be endorsed for elephant management in open contexts.

Large carnivores inhabiting ecosystems with heterogeneously distributed environmental resources with strong seasonal variations frequently employ opportunistic foraging strategies, often typified by seasonal switches in diet. In semi-arid ecosystems, herbivore distribution is generally more homogeneous in the wet season, when surface water is abundant, than in the dry season when only permanent sources remain. Here, we investigate the seasonal contribution of the different herbivore species, prey preference and distribution of kills (i.e. feeding locations) of African lions in Hwange National Park, Zimbabwe, a semi-arid African savanna structured by artificial waterholes. We used data from 245 kills and 74 faecal samples. Buffalo consistently emerged as the most frequently utilised prey in all seasons by both male (56%) and female (33%) lions, contributing the most to lion dietary biomass. Jacobs’ index also revealed that buffalo was the most intensively selected species throughout the year. For female lions, kudu and to a lesser extent the group “medium Bovidae” are the most important secondary prey. This study revealed seasonal patterns in secondary prey consumption by female lions partly based on prey ecology with browsers, such as giraffe and kudu, mainly consumed in the early dry season, and grazers, such as zebra and suids, contributing more to female diet in the late dry season. Further, it revealed the opportunistic hunting behaviour of lions for prey as diverse as elephants and mice, with elephants taken mostly as juveniles at the end of the dry season during droughts. Jacobs’ index finally revealed a very strong preference for kills within 2 km from a waterhole for all prey species, except small antelopes, in all seasons. This suggested that surface-water resources form passive traps and contribute to the structuring of lion foraging behaviour.

No abstract